Hi Julia, > -----Original Message----- > From: McWhirter,Julia > > Larry/Jennifer, > > I have copied the chickenpox and popcorn rules which are > working fine thanks very much Jennifer. Larry my install is > also outside the firewall and therefore needs site-wide > config and not user based and therefore getting bayes working > is going to mean collecting a lot of hams/spams and > transferring around (bit of a pain) but currently I have just > started collecting to one day get it to learn so that in the > future the need for some rules will be negated as you say.
Just thought I would add my experience in the outside chance it could help. I am using Procmail in my site-wide config. I was also utilizing Procmail for whitelist. What I decided to do is turn off my whitelist and instead force feed those messages to sa-learn as ham *just* prior to feeding to spamc. I have been utilizing AWL along with auto_learn. To date, it has been working great. About 99% of the legitimate newsletters requested advertisements that used to be FPs are not being tagged as spam anymore but the spam hit is actually more accurate. I have only had to hand-feed maybe 15-20 messages as either ham or spam to sa-learn. It has mostly been hands-off. The only thing I really spent any time with is removing the newsletters and requested advertisements from AWL so that it would rescore. I am not disappointed. --Larry ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk